Results

Cardiovascular reactivity (especially SBP) was the primary dependent variable. In the case of significant associations between respective baseline and reactivity values, the reactivity scores were corrected using the baseline as a covariate to prevent carry-over or initial values effects (Llabre, Spitzer, Saab, Ironson, & Schneiderman, 1991). The hypotheses were tested with 2 (Mood) × 2 (Memorizing) × 2 (Time) mixed model ANOVAs with repeated measures on the last factor. In the case of significant interactions including the time factor, we ran separate 2 x 2 between persons ANOVAs of cardiovascular reactivity during the mood inductions and task performance. Given the directed nature of our theory-based predictions, follow-up comparisons between the experimental conditions were made with one-tailed cell contrasts. Task performance and verbal measures were analyzed with 2 (Mood) × 2 (Memorizing) between-persons ANOVAs.

Mood manipulation check

We created mood sum scores for the mood baseline (Cronbach’s α = .68) and post-mood manipulation (Cronbach’s α = .82) measures by reversing the negative hedonic tone items and adding them to the positive hedonic tone items. A 2 (Mood) x 2 (Memorizing) between-persons ANOVA of the mood baseline score found no significant effects (all ps > .07; average M = 19.31, SE = 0.56). We calculated mood change scores by subtracting the mood baseline scores from the post-manipulation scores. A 2 x 2 between-persons ANCOVA of these reactivity scores, using the mood baseline scores as covariate, found a significant covariate effect, F(1, 44) = 12.74, p < .002, and, most relevant, a strong Mood main effect, F(1, 44) = 17.71, p < .001, indicating elated mood in the positive mood condition (M = 1.29, SE = 0.67) and depressed mood in the negative mood condition (M = -2.70, SE = 0.65).

Cardiovascular measures: Preliminary analysis

A preliminary analysis revealed gender main effects on the baseline values of SBP, DBP, and HR, Fs(1, 42) > 6.65, ps < .01. Men had higher blood pressure and lower heart rate (SBP: M = 124.60, SE = 15.86;

DBP: M = 71.12, SE = 10.81; HR: M = 66.11, SE = 8.79) than women (SBP: M = 110.63, SE = 11.83; DBP: M

= 61.38, SE = 7.04; HR: M = 75.33, SE = 10.69), which is a typical finding.

Cardiovascular baselines

The averages of the cardiovascular measures of the last two minutes of the habituation period constituted the baseline values for SBP, DBP, and HR (Cronbach’s αs > .97)⁷. Cell means and standard errors appear in Table 1. According to 2 x 2 between-persons ANOVAs, there were no significant baseline differences between conditions for any measure (all ps > .05).

Table 1. Cell Means and Standard Errors (in Parentheses) of the Cardiovascular Baseline Values.

Intentional-memorizing Incidental-memorizing

Negative mood Positive mood Negative mood Positive mood

SBP 112.79 126.45 114.98 112.26

(3.50) (5.02) (4.24) (3.52)

DBP 63.15 70.85 65.60 62.53

(2.53) (3.36) (2.98) (2.01)

HR 72.04 71.30 71.41 71.11

(3.62) (3.08) (3.86) (1.71)

Note: Cell ns = 12 for intentional-memorizing conditions and cell ns = 13 for incidental-memorizing conditions; SBP:

systolic blood pressure; DBP: diastolic blood pressure; HR: heart rate. Units of measure are millimeters of mercury for SBP and DBP, beats per minute for HR.

Cardiovascular reactivity

Change scores were computed for each participant by subtracting the baseline values from the averages of values obtained during the mood induction and the task performance periods (Cronbach’s αs > .98).

SBP reactivity. Cell means are depicted in Figure 6 (top panel). A 2 (Mood) x (Memorizing) x 2 (Time) mixed model ANCOVA, using the SBP baselines as covariate, found a reliable Time x Baseline interaction, F(1, 45) = 7.33, p < .01, reflecting that SBP reactivity was associated with the baselines during the mood

7 We calculated the physiological baseline values from the last two minutes of the habituation period because for all parameters the last two minutes did not differ significantly from one another (ps > .31).

inductions but not during task performance (see below). The analysis also revealed a Memorizing x Time interaction, F(1, 45) = 27.37, p < .001, which was qualified by the expected three-way interaction between Mood, Memorizing, and Time, F(1, 45) = 5.37, p < .03. No other effect was significant (ps > .10).

A separate between-persons ANCOVA of SBP reactivity during the mood inductions found only a significant covariate (baseline) effect, F(1, 45) = 15.32, p < .001, but no other reliable effects (ps > .20).

For the task performance period the covariate effect was not significant (p > .07). Here, an ANOVA revealed a Memorizing main effect, F(1, 46) = 16.53, p < .001, reflecting stronger reactivity in the intentional-memorizing condition (M = 5.84, SE = 0.90 vs. M = 0.75, SE = 0.87), and, most relevant, a significant Mood x Memorizing interaction, F(1, 46) = 5.10, p < .03. As predicted, cell contrasts found that SBP reactivity in the intentional-memorizing condition was significantly stronger in a negative mood (M = 8.33, SE = 1.15), than in a positive mood (M = 3.36, SE = 2.01), t(46) = 2.14, p < .02. By contrast, the mood effect in the incidental-memorizing condition was not significant (negative mood: M = 0.40, SE = 0.70; positive mood: M = 1.09, SE = 0.85; p > .50).

DBP reactivity. As depicted in Figure 6 (bottom panel), diastolic reactivity resembled the SBP responses. The analysis revealed again a significant Time x Covariate interaction, F(1, 45) = 5.73, p < .02, indicating that the strength of association between baselines and reactivity scores differed between the two measurement periods (see below). Moreover, there was a significant Memorizing x Time interaction, F(1, 45) = 24.74, p < .001, that was further qualified by a reliable three-way interaction, F(1, 45) = 4.54, p < .04. The ANCOVA of DBP reactivity during the mood inductions found no significant effects (ps > .35), except for a highly significant covariate effect, F(1, 45) = 13.36, p < .001. For the task performance period, the covariate was not significant, F(1, 45) = 3.74, p > .06, but there were a reliable Memorizing main effect, F(1, 46) = 22.51, p < .001, due to stronger reactivity in the intentional-memorizing condition (M = 3.88, SE = 0.58 vs. M = 0.09, SE = 0.55), and a marginally significant Mood x Memorizing interaction, F(1, 46) = 3.86, p < .06. Additional cell comparisons found for the intentional-memorizing condition significantly stronger DBP reactivity in a negative mood (M = 5.08, SE = 0.95) than in a positive mood (M = 2.68, SE = 1.12), t(46) = 2.08, p < .03. By contrast, the mood effect in the incidental-memorizing condition was not significant (negative mood: M = -0.28, SE = 0.59; positive mood: M = 0.46, SE = 0.22; p > .25).

A: Systolic blood pressure (SBP) reactivity

HR reactivity. Since an ANCOVA did not find any significant baseline effects (ps > .37), reactivity values were not baseline-corrected. The 2 x 2 x 2 ANOVA revealed a strong Time main effect, F(1, 46) = 98.88, p < .001, reflecting stronger reactivity during task performance (M = 1.95, SE = 0.71) than during the mood inductions (M = -3.44, SE = 0.48). Moreover, there was a marginally significant Time x Memorizing interaction, F(1, 46) = 3.84, p < .06. There was no memorizing effect during the mood inductions, t(48) = 1.21, p > .23., but during task performance, t(48) = 2.31, p < .03, indicating weaker reactivity in the incidental-memorizing condition than in the intentional-memorizing condition (M = 0.30, SE = .97 vs. M = 3.61, SE = 1.06).

Task performance

A 2 (Mood) × 2 (Memorizing) between-persons ANOVA of the total number of recalled letter series revealed only a significant Memorizing main effect, F(1, 45) = 5.73, p < .05, reflecting more recalled series in the intentional than in the incidental memorizing condition (M = 6.91, SE = 0.50 vs. M = 5.27, SE

= 0.47). An analysis of the number of correctly recalled items revealed the same main effect, F(1, 45) = 11.19, p < .003, (M = 5.18, SE = 0.55 vs. M = 2.65, SE = 0.52). No further effects attained significance (all ps >.27).